[PATCH v2 net-next] be2net: disable bh with spin_lock in be_process_mcc

[PATCH v2 net-next] be2net: disable bh with spin_lock in be_process_mcc

[Denis Kirjanov]

be_process_mcc() is invoked in 3 different places and
always with BHs disabled except the be_poll function
but since it's invoked from softirq with BHs
disabled it won't hurt.

v1->v2: added explanation to the patch

Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>
---
 drivers/net/ethernet/emulex/benet/be_cmds.c | 4 ++--
 drivers/net/ethernet/emulex/benet/be_main.c | 2 --
 2 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/emulex/benet/be_cmds.c b/drivers/net/ethernet/emulex/benet/be_cmds.c
index ef5d61d57597..9365218f4d3b 100644
--- a/drivers/net/ethernet/emulex/benet/be_cmds.c
+++ b/drivers/net/ethernet/emulex/benet/be_cmds.c
@@ -550,7 +550,7 @@ int be_process_mcc(struct be_adapter *adapter)
 	int num = 0, status = 0;
 	struct be_mcc_obj *mcc_obj = &adapter->mcc_obj;
 
-	spin_lock(&adapter->mcc_cq_lock);
+	spin_lock_bh(&adapter->mcc_cq_lock);
 
 	while ((compl = be_mcc_compl_get(adapter))) {
 		if (compl->flags & CQE_FLAGS_ASYNC_MASK) {
@@ -566,7 +566,7 @@ int be_process_mcc(struct be_adapter *adapter)
 	if (num)
 		be_cq_notify(adapter, mcc_obj->cq.id, mcc_obj->rearm_cq, num);
 
-	spin_unlock(&adapter->mcc_cq_lock);
+	spin_unlock_bh(&adapter->mcc_cq_lock);
 	return status;
 }
 
diff --git a/drivers/net/ethernet/emulex/benet/be_main.c b/drivers/net/ethernet/emulex/benet/be_main.c
index 2edb86ec9fe9..4d8e40ac66d2 100644
--- a/drivers/net/ethernet/emulex/benet/be_main.c
+++ b/drivers/net/ethernet/emulex/benet/be_main.c
@@ -5630,9 +5630,7 @@ static void be_worker(struct work_struct *work)
 	 * mcc completions
 	 */
 	if (!netif_running(adapter->netdev)) {
-		local_bh_disable();
 		be_process_mcc(adapter);
-		local_bh_enable();
 		goto reschedule;
 	}
 
-- 
2.12.3

Re: [PATCH v2 net-next] be2net: disable bh with spin_lock in be_process_mcc

[Willem de Bruijn]

On Thu, Aug 1, 2019 at 5:24 AM Denis Kirjanov <kda@linux-powerpc.org> wrote:
>
> be_process_mcc() is invoked in 3 different places and
> always with BHs disabled except the be_poll function
> but since it's invoked from softirq with BHs
> disabled it won't hurt.

This describes the current state. What is the benefit of removing the
local_bh_disable/local_bh_enable pair from one caller (be_worker), but
not another (be_mcc_wait_compl) and then convert process_mcc to
disable bh itself with spin_lock_bh?

Re: [PATCH v2 net-next] be2net: disable bh with spin_lock in be_process_mcc

[David Miller]

From: Denis Kirjanov <kda@linux-powerpc.org>
Date: Thu,  1 Aug 2019 11:24:20 +0200

> be_process_mcc() is invoked in 3 different places and
> always with BHs disabled except the be_poll function
> but since it's invoked from softirq with BHs
> disabled it won't hurt.
> 
> v1->v2: added explanation to the patch
> 
> Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>

Like Willem I see no benefit at all to this change.